How to do the plot of standardized deviance residual against the predicted value of the index function after fitting logit model in Stata?

[smith Jason]

As suggested by the title, I want to fit a logit model and graph the plot of the standardized deviance residual against the predicted value of the index function.
However, I really don't know how to do this in Stata. Meanwhile, I don't know how to graph the residual-predictor plot in Stata as well.
Can someone help me?
Thank you!

webuse auto,clear
logit foreign price mpg rep78

[Carlo Lazzaro]

Jason:
see -help regress postestimation plots-.

[smith Jason]

Jason:
see -help regress postestimation plots-.

I used rvpplot command to do it and it doesn't work at all.
Anyway, Thank you!

[Carlo Lazzaro]

Jason:
do you mean something along the following lines?

. use https://www.stata-press.com/data/r17/auto.dta
(1978 automobile data)

. logit foreign i.rep78 mpg

<snip>

Logistic regression                                     Number of obs =     59
                                                        LR chi2(3)    =  25.87
                                                        Prob > chi2   = 0.0000
Log likelihood = -25.478287                             Pseudo R2     = 0.3367

------------------------------------------------------------------------------
     foreign | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
       rep78 |
          1  |          0  (empty)
          2  |          0  (empty)
          3  |  -3.136422   1.044601    -3.00   0.003    -5.183803    -1.08904
          4  |  -1.119903   .9741478    -1.15   0.250    -3.029198    .7893916
          5  |          0  (omitted)
             |
         mpg |   .1310946    .070733     1.85   0.064    -.0075396    .2697287
       _cons |  -1.723275   1.776453    -0.97   0.332    -5.205059    1.758509
------------------------------------------------------------------------------

. predict fitted, xb


. predict dev_res, deviance


. twoway (scatter dev_res fitted)

.

[Carlo Lazzaro]

Jason:
if what above does not work, please share an example/experpt of your dataset via -dataex-. Thanks.

[smith Jason]

I just asked my friend and he told me that the code should be as follows,

webuse auto,clear
logit foreign price mpg rep78
rvpplot2 weight, yli(0)

[smith Jason]

Jason:
if what above does not work, please share an example/experpt of your dataset via -dataex-. Thanks.

Thank you!
But I really don't know if what you just posted is the solution to my question.

[smith Jason]

Jason:
do you mean something along the following lines?

. use https://www.stata-press.com/data/r17/auto.dta
(1978 automobile data)

. logit foreign i.rep78 mpg

<snip>

Logistic regression Number of obs = 59
LR chi2(3) = 25.87
Prob > chi2 = 0.0000
Log likelihood = -25.478287 Pseudo R2 = 0.3367

------------------------------------------------------------------------------
foreign | Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
rep78 |
1 | 0 (empty)
2 | 0 (empty)
3 | -3.136422 1.044601 -3.00 0.003 -5.183803 -1.08904
4 | -1.119903 .9741478 -1.15 0.250 -3.029198 .7893916
5 | 0 (omitted)
|
mpg | .1310946 .070733 1.85 0.064 -.0075396 .2697287
_cons | -1.723275 1.776453 -0.97 0.332 -5.205059 1.758509
------------------------------------------------------------------------------

. predict fitted, xb


. predict dev_res, deviance


. twoway (scatter dev_res fitted)

.

Thank you!
But I really don't know if what you just posted is the solution to my question.
predicted value of the index function====predict fitted, xb?

[Carlo Lazzaro]

Jason:

webuse auto,clear
logit foreign price mpg rep78
rvpplot2 weight, yli(0)

how can it be? -weight- is not included as a predictor in the right-hand side of your regression equation.
That said:
1) the index function details the relationship between the covariates (right-hand side of your regression equation) and the outcome (left-hand side of your regression equation) (see, if interested,

https://www.stata.com/bookstore/heal...cs-using-stata, page 73).

2) if you have a different notion/idea/whatever about 1), please share it with the list ( the risk of an xy problem cannot be ruled out on a priori basis; see, if interested: https://xyproblem.info ;
3) as per FAQ, stating that a given command/function/whatever does not work as expected, is not informative.

[smith Jason]

Thank you!
It is a typo. It should be price rather than weight.